chapter 2 achievement of a synergistic adjuvant effect on ... · chapter 2 achievement of a...

Chapter 2

Achievement of a synergistic adjuvant effect on arthritis induction by activation of innate immunity and forcing the immune response toward the

Th1 phenotype

Anita Hanyecz*1, Suzanne E. Berlo*1, Sándor Szántó1, Chris P.M. Broeren2†, Katalin Mikecz1 and Tibor T. Glant1

*Authors were equally involved in this work

1Section of Molecular Medicine, Department of Orthopedic Surgery, Rush University Medical Center, Chicago, IL 60612. 2Department of Infectious Diseases and Immunology, Division of Immunology, Faculty of Veterinary Medicine, Utrecht University, Utrecht, The Netherlands. Arthritis and Rheumatism 2004; 50(5):1665-1676

Chapter 2

32

Abstract Objective. To apply and analyze the mechanisms of action of dimethyldioctadecylammonium bromide (DDA), a powerful adjuvant that does not have the side effects of conventionally used complete Freund’s adjuvants in proteoglycan-induced arthritis (PGIA) and collagen-induced arthritis (CIA). Methods. PGIA and CIA were generated using standard immunization protocols with cartilage proteoglycan (PG) aggrecan or human type II collagen (CII) emulsified with CFA, and compared with PGIA and CIA generated using immunization protocols in which the same antigens were used in combination with the adjuvant DDA. Immune responses to immunizing and self PGs and CII, and the incidence, severity, and onset of arthritis were monitored throughout the experiments. In addition, a new, inexpensive, and powerful method of inducing arthritis using crude cartilage extracts is described. Results. A significantly reduced onset period and a more severe arthritis were achieved in BALB/c mice immunized with cartilage PGs in DDA. PGs from bovine, ovine, and porcine cartilage, which otherwise have no effect or have only a subarthritogenic effect, and crude extracts of human osteoarthritic cartilage induced a 100% incidence with a very high arthritis score in BALB/c mice. The overall immune responses to either CII or PG were similar in antigen/CFA-immunized and antigen/DDA-immunized animals, but the Th1/Th2 balance shifted significantly toward a Th1 bias in DDA injected animals with either PGIA or CIA. Conclusion. DDA, which was first used in autoimmune models, is a potent nonirritant adjuvant, which eliminates all undesired side effects of the Freund’s adjuvants. DDA exerts a strong stimulatory effect via the activation of nonspecific (innate) immunity and forces the immune regulation toward Th1 dominance. These lines of evidence also suggest the possibility that seemingly innocuous compounds may exert an adjuvant effect in humans and may create the pathophysiologic basis of autoimmunity in susceptible individuals via the activation/stimulation of innate immunity.

DDA replaces Freund’s adjuvants in PG-induced arthritis

33

Introduction Several lines of evidence indicate that the effector mechanism that initially attacks the small joints in rheumatoid arthritis (RA) is T cell driven. As a result, an aggressive synovial pannus develops that destroys articular cartilage and bone, leading to massive ankylosis and deformities of the peripheral joints. The disease has a progressive character, with involvement of more and more joints over time. While the primary target organ is the synovial joint, there is no clear evidence that any macromolecule of cartilaginous tissues, bone, or synovium would be a preferential autoantigen. Nevertheless, the most relevant animal models of RA appear to be those induced by cartilage matrix components, such as type II collagen (CII) or proteoglycan aggrecan (PG). Systemic immunization of genetically susceptible rodents with cartilage-specific CII (1,2) or of BALB/c or C3H mice

with human cartilage PG (3-7) leads to the development of progressive polyarthritis (8). Both models use 100-150 µg

of antigenic protein (either CII or cartilage PG) in Freund’s complete adjuvant (CFA). Mycobacterial heat-shock proteins (HSPs) present in CFA are known to be very potent nonspecific immunostimulators, and these proteins significantly contribute to, and potentiate, the systemic T cell response (9,10). In subsequent injections, the antigen can be administered in emulsion with mineral oils (Freund’s incomplete adjuvant; IFA) that do not contain mycobacterial components. Mineral oils or oil derivatives, such as pristane, can also provoke inflammatory reactions in the synovial joints of certain (susceptible) rodent strains (11-13). Due to an unknown mechanism, peripheral joints are the target organs in RA, and a number of animal models can simulate the human disease, but all require antigen challenge with adjuvant (14,15). In an attempt to avoid the effect of bacterial HSPs in autoimmune models of arthritis, as well as CFA-induced irritation and granuloma formation with subsequent adhesions in the peritoneal cavity (15), we used a positively charged lipophilic quaternary amine (dimethyldioctadecylammonium bromide; DDA) instead of CFA. With the use of DDA as adjuvant, the time of onset of arthritis was significantly earlier and the severity was significantly greater in PG-induced arthritis (PGIA). In collagen-induced arthritis (CIA), a significantly higher incidence was achieved. The combination of the positively charged lipophilic adjuvant DDA (C38H80NBr) with the negatively charged cartilage PG seems to be powerful and simultaneously avoids the detrimental and destructive effects of CFA. We also provide herein a simple and convenient protocol for the preparation of crude cartilage extract that can be used to induce PGIA in BALB/c mice with maximum incidence (100%) and severity. The use of crude cartilage extract with DDA does not require the purification of PG by cesium chloride gradient ultracentrifugation, and PGs from human (osteoarthritic), canine, porcine, and bovine cartilage all proved equally arthritogenic in genetically susceptible BALB/c mice.

Chapter 2

34

Materials and Methods The use of human materials was approved by the Institutional Review Board, and the use of animals for immunization and arthritis induction was approved by the Institutional Animal Care and Use Committee, of Rush University. Isolation of antigenic components from cartilage Crude (total) cartilage extracts. Total cartilage extracts were obtained by 4M guanidinium chloride extraction (6,16) from newborn and adult human and bovine articular cartilage, bovine nasal cartilage, ewe and swine articular cartilage, chicken sternal cartilage, cartilaginous skeletal tissue from newborn mice, and rat chondrosarcoma. These cartilage extracts were further purified by gradient centrifugation to obtain highly purified cartilage PGs or were dialyzed against water and lyophilized (hereafter called crude cartilage extract). High-density cartilage PG. High-density cartilage PG was purified by repeated CsCl gradient ultracentrifugation under dissociative conditions (in 4M guanidinium chloride), as described in detail elsewhere (6,16). Purified cartilage PGs were deglycosylated for immunization or were left untreated. None of the purified PGs (“native PG”) or the crude cartilage extracts without deglycosylation induced arthritis, and the results with deglycosylated cartilage samples are therefore presented. Purified PGs or crude extracts of normal or osteoarthritic human cartilage were treated with chondroitinase ABC (Seikagaku America/Associates of Cape Cod, Falmouth, MA), 5 units/100 mg of PG or 5 units/1 g of crude cartilage extract in 0.1M Tris–acetate buffer, pH 7.6, for 24 h at 37°C, to deplete chondroitin sulfate side chains (3,6,16). Alternatively, we achieved the same level of chondroitin sulfate depletion by digesting samples with testicular hyaluronidase (Worthington Biochemical, Lakewood, NJ), using 5,000 units/100 mg of purified PG or 5,000 units/g of crude cartilage extract in sodium acetate buffer (pH 5.0) containing 50 mM NaCl and 50 mM Mg2SO4 (24 h at 37°C) (5). PG samples or crude extracts of cartilage isolated from skeletally mature adults of different species (Table 1) were

subsequently digested with endo-β-galactosidase (Seikagaku America), 0.1 unit/100 mg of PG or 0.1 unit/1 g of

crude cartilage extract in sodium acetate buffer, pH 5.8, to remove residual keratan sulfate side chains present in adult (aging) cartilage PGs (5,6,16). Since the keratan sulfate and chondroitin sulfate side chains may mask a number of dominant/arthritogenic epitopes, the depletion of these glycosaminoglycan side chains is critical in order to “retrieve” dominant arthritogenic T cell epitopes of the core proteins of PG (4,5,8). A PG/glycosaminoglycan-free crude cartilage extract was obtained by diethylaminoethyl (DEAE; Whatman, Clifton, NJ) ion-exchange chromatography, as described previously (17). The unbound fraction was retrieved at 0.15M NaCl and was further absorbed with hyaluronan-Sepharose (18) to remove glycosaminoglycan-free G1-domain fragments of PG and cartilage link protein (5). Samples were dialyzed against water and lyophilized. The absence of glycosaminoglycans in the PG-free fractions of crude extracts of human cartilage (Table 1) was confirmed by monoclonal antibodies (mAb) to chondroitin 4-sulfate (clone BE123), chondroitin 6-sulfate (clone MK302), and keratan sulfate (clone EFG11) (Chemicon, Temecula, CA). The absence of the G1 domain was confirmed by


35

mAb G18 (5) and the absence of the link protein by a rabbit polyclonal antibody LP1-R18 raised against a link protein-specific synthetic peptide (19). The DEAE-bound fraction was eluted with 2.0M NaCl, deglycosylated with

chondroitinase ABC and endo-β-galactosidase (human adult articular cartilage PG + ABC/endo-β-galactosidase) as

described above, and used as a positive control for immunization (Table 1). Human cartilage residues remaining after guanidinium chloride extraction were washed exhaustively with cold water, equilibrated in 0.5M acetic acid, and then digested with pepsin (1 mg of enzyme/g wet weight of cartilage residue; Worthington) at 4°C as described previously (6,20,21). Type II collagen was purified by sequential and repeated precipitation with NaCl in 0.5M acetic acid at 4°C. Animals and immunization protocols More than 1,200 mice of different strains, ages, and both sexes were purchased from the National Cancer Institute (NCI; Frederick, MD), Charles River Laboratories (Kingston, NY, Raleigh, NY, and Portage, MI colonies), Harlan (Madison, WI), Taconic Farm (Germantown, NY) and The Jackson Laboratory (Bar Harbor, ME). Female SCID mice of BALB/c background (NCI/NCrC.B-17-scid/scid) were obtained from the NCI and maintained under germ-free conditions; these mice were used for transfer experiments. Female and male mice (12-16 weeks of age) of different inbred strains as well as their F1 hybrids were used to compare selected immunization protocols. In experiments in which the effects of different cartilage PGs and/or crude cartilage extracts, with or without treatment of different glycosidases, were compared (Table 1 and Fig. 1), we immunized female retired breeder BALB/c mice that had been purchased from the NCI. All mice were housed in the same room of the Comparative Research Center at Rush University. When a new experimental group was immunized, 10-15 female retired breeder BALB/c mice were also immunized with deglycosylated human cartilage PGs in DDA or CFA; these groups were used as positive/reference control groups.

Mice were injected intraperitoneally with 100 µg of deglycosylated cartilage PGs or crude extracts (both measured

and normalized to PG core protein) (6) or with human cartilage CII, except where indicated otherwise. A fine cationic liposome form of DDA (micelle) was obtained by heating a 10-mg/ml DDA suspension (Sigma-Aldrich, St. Louis, MO) in phosphate buffered saline (PBS; pH 7.4) to 56-63°C for 15-20 minutes and then cooling on ice (22). Alternatively, the same effect of micelle formation can be achieved by heating the DDA in a microwave oven. This DDA micelle form was mixed with an equal volume of antigen (1 mg of CII/ml or 1 mg of PG core protein/ml of PBS), and the mixture was either shaken for 20-30 minutes at room temperature or vortexed for 15-20 seconds. DDA and antigen/DDA (CII/DDA or PG/DDA) emulsions were freshly prepared on the day of immunization and stored on ice

until the time of injection. The antigen/DDA micelle emulsion (200 µl total) was injected intraperitoneally (IP) or,

where indicated, subcutaneously (SC) or intradermally (ID). Four major groups of animals were treated according to 1 of the following “standard” immunization protocols. In

protocol 1 (PG/CFA), an IP injection of 100 µg of PG protein in 100 µl of PBS (pH 7.4) emulsified with 100 µl of CFA

(Difco, Detroit, MI) was given on day 0. On days 21 and 42, the same dose of PG in IFA (Difco) was injected IP. In

protocol 2, the same antigen dose, injection time points, and IP approach were used, but the 100 µg of PG in 100 µl

Chapter 2

36

of PBS was mixed with 1 mg of DDA micelle in 100 µl of PBS. In protocol 3, 100 µg of CII was dissolved in 0.1M

acetic acid, and the volume was adjusted to 1 mg of CII/ml of PBS to prepare an emulsion with an equal volume of

CFA. The emulsion containing 100 µg of CII was then injected either IP or ID into the proximal tail on days 0 and 21.

In protocol 4, the same dose of CII (100 µg) in 100 µl of DDA micelle (1 mg) was also used for either IP or ID

immunization. In addition to these 4 “standard” immunization protocols, we used 5 other approaches. In protocols 5 and 6, different doses (protocol 5) of PGs or crude cartilage extracts (protocol 6) were mixed with different amounts of DDA and injected IP, SC, or ID. A minimum of 0.5 mg of DDA per injection was required to achieve the maximum effect (incidence and severity) in BALB/c mice, but the onset of arthritis was faster when 1 mg of DDA micelle was used. Higher doses did not change any immune or clinical parameters of arthritis induction, and the IP immunization was the most effective for inducing arthritis. Concentrations higher than 20-25 mg/ml of DDA micelle were insoluble, and

3.5-4.0 mg of DDA in 100 µl per mouse per injection in 100 µl resulted in 15-20% mortality within 4-5 days after IP

injection. In protocol 7, we also used PG/DDA in micelle form (heated and cooled as described above), or the water-insoluble DDA was simply suspended in PBS at room temperature and mixed with PG. Although the PG/DDA suspension produced a slightly higher Th1 response and the arthritis onset was 1-3 days earlier than when PG/DDA micelle was used (data not shown), the PG/DDA suspension required constant shaking, which made the immunization more

complicated and less reproducible. In protocol 8, we also used purified cartilage PG (100 µg of core protein) in

emulsion with DDA micelle (1 mg) for immunization as a standard method, and we compared the results with those obtained when the PG and DDA were injected separately into different areas (e.g., DDA injected IP and PG injected SC or vice versa). IP administration of DDA appeared to be critical to achieving a maximum adjuvant effect in PGIA, and ID administration appeared to be critical in CIA. In contrast, IP injection of DDA was not an irritant, and no

granuloma formation or adhesion occurred. Therefore, unless indicated otherwise, we routinely used 100 µg of PG

core protein in 100 µl of PBS with 1 mg of DDA micelle in 100 µl of PBS for IP immunization of BALB/c mice

purchased from NCI. Along with these preliminary/supplemental experiments, we also immunized BALB/c mice with keyhole limpet

hemocyanin, ovalbumin, bovine serum albumin (BSA), or methylated BSA (mBSA) (100 µg of antigen/injection; all

from Sigma-Aldrich) in DDA (protocol 9). Arthritis was scored, and T and B cell responses and cytokine levels were measured as described below for PG/DDA-immunized animals. Assessment of arthritis The paws of all immunized mice were examined twice weekly until day 21, and then daily thereafter. Abnormalities due to arthritic changes of the joints were recorded. The appearance of joint swelling was recorded as the time of onset of arthritis. A standard scoring system based upon swelling and redness (range 0-4 for each paw; maximum possible score 16 per animal) was used for the assessment of disease severity (3,8,21,23). The limbs and spine of


37

arthritic and nonarthritic mice were dissected, fixed, decalcified, sectioned, and the tissue sections were stained with hematoxylin and eosin for histopathologic examination. Measurement of antigen-specific T cell responses, antibodies, and cytokines Sera were collected from immunized mice during the immunization period (once or twice each week from the retroorbital venous plexus) and at the end of the experiments. Spleen and lymph node cells were collected at the end of the experiments. PG-specific antibodies were measured by enzyme-linked immunosorbent assay (ELISA) as described previously (7,23,24). Maxisorp immunoplates (Nunc, Roskilde, Denmark) were coated with human or

mouse cartilage PGs (0.1 µg of protein/well), and the free binding sites were blocked with 1% fat-free milk in PBS.

Sera were applied at increasing dilutions, and levels of both total anti-PG antibodies and isotypes of PG-specific antibodies were determined using peroxidase-conjugated goat anti-mouse IgG (Accurate, Westbury, NY) or rat mAb to mouse IgG1 or IgG2a (Zymed, South San Francisco, CA) as secondary antibodies (23). Serum antibody levels were calculated relative to the corresponding mouse IgG isotype standards (all from Zymed) or mouse serum immunoglobulin fractions (Sigma-Aldrich) (23,24). The total serum IgG fraction was determined by a capture ELISA

method (23) using mouse κ-chain-specific peroxidase-labeled rat mAb for detection (BioSource, Camarillo, CA).

Antigen-specific T cell responses were measured in quadruplicate samples of spleen cells or PG/adjuvant-primed

lymph node cells (3 x 105 cells/well) cultured in the presence of 25 µg of PG protein/ml or 25-50 µg of CII. Antigen-

specific interleukin-2 (IL-2) production was measured in 2-day-old supernatant by CTLL-2 bioassay, and T cell proliferation was assessed on day 5 by 3H-thymidine incorporation (5-7,23,24). The antigen-specific T cell response was expressed as a stimulation index (S.I.), which was calculated as the ratio of the counts per minute of 3H-thymidine incorporated in antigen-stimulated cultures relative to that in nonstimulated cultures. Production of antigen-

specific interferon-γ (IFNγ), IL-4, IL-12, and tumor necrosis factor α (TNFα) was measured in cell culture

supernatants (3 x 106 cell/ml) on day 4 using a capture ELISA method (BD Biosciences, San Diego, CA, or R&D

Systems, Minneapolis, MN) as described previously (7,24). Cytokines (IFNγ, IL-1α, IL-6, IL-10, IL-12, and TNFα) in

the sera of immunized animals were measured by ELISA at the end of the experiments. Characterization of cell surface markers The expression of cell surface markers was assayed by flow cytometry. Briefly, either 1 x 106 unseparated spleen

cells obtained at different time points of immunization, lymph node cells from antigen-primed mice (100 µg of antigen

with 100 µl of CFA or DDA micelle were injected into the footpad and cells harvested 9 days later), or cells harvested

from peritoneal lavage fluid were incubated with fluorescein isothiocyanate- or phycoerythrin-labeled or biotinylated mAb to CD45/B220 (B cell marker), CD3, CD4, CD8, CD25, CD28, CD44, and CD69 (T cell and T cell activation markers), mAb to CD68 or F4/80 (monocyte/macrophage marker), mAb to Gr-1 (myeloid cell lineage marker, e.g., neutrophil leukocytes), and mAb to CD11c (activated dendritic cell marker). Antibodies were purchased from BD Biosciences, R&D Systems, or BioSource.

Chapter 2

38

Cells were washed twice in PBS containing 1% BSA and 0.02% sodium azide. The cells labeled with biotinylated primary antibodies were incubated with streptavidin-R-phycoerythrin (BD Biosciences) or streptavidin-Alexa Fluor 488 (Molecular Probes, Eugene, OR) for an additional 30 minutes at 4°C. After 2 more washes, the cells were fixed in 2% buffered formalin, and samples were analyzed on a FACScan flow cytometer (Becton Dickinson, San Jose, CA) using CellQuest software (Becton Dickinson). Cell isolation and transfer of arthritis Single-cell suspensions were prepared from the spleens of arthritic BALB/c mice. Donor arthritic mice were immunized with cartilage PG in CFA or in DDA as described above for protocols 1 and 2. Mononuclear cells were isolated on Lympholyte M (Zymed) and used for adoptive transfer of arthritis as described previously (23). In all

transfer experiments, 1.5 x 107 unseparated spleen cells were injected IP on day 0 with 100 µg of cartilage PG

(measured as protein), and 1 x 107 spleen cells from arthritic donors were injected on day 7 without cartilage PG into SCID mice. There were absolutely no differences in the incidence, onset, severity, or histopathologic features of adoptively transferred arthritis in SCID mice that received donor cells from PG/CFA-immunized (23) or PG/DDA-immunized (data not shown) mice. Statistical analysis Statistical analysis was performed using SPSS version 7.5 software (SPSS, Chicago, IL). The Mann-Whitney and Wilcoxon tests were used for intergroup comparisons. Both the 5% significance level and the 1% significance level were used.


39

Results Incidence and severity of PGIA in arthritis-susceptible BALB/c and C3H colonies DDA was first used in our laboratory when we studied how HSPs affect the onset and severity of PGIA (25), since we wanted to avoid the undesired effects of the bacterial HSPs present in CFA. A group that received PG with DDA (PG/DDA) IP developed more severe PGIA significantly earlier than did BALB/c mice immunized with the same dose of PG with CFA and IFA (Fig. 1).

A BA B

Figure 1. Incidence and severity of proteoglycan-induced arthritis (PGIA) in BALB/c mice immunized with human cartilage proteoglycan aggrecan (PG) in Freund’s complete adjuvant (CFA), Freund’s incomplete adjuvant (IFA), or dimethyldioctadecylammonium bromide (DDA). Human PG was deglycosylated of both chondroitin sulfate and keratan sulfate side chains (see Materials and Methods). A, The incidence of arthritis was assessed twice a week. B, Each paw of each animal was scored for inflammation (0-4 scale) resulting in an arthritis score for each immunized animal (range of possible scores 0-16). Arrows mark the injection times for PG/CFA (solid arrows), PG/IFA (shaded arrows), and PG/DDA (open arrows with solid arrowheads). The cumulative results of 4 independent experiments (n = 10-15 animals per experiment) are shown. Since PGIA can be induced only in genetically susceptible BALB/c and C3H strains, but the severity and susceptibility are different in different colonies of the same murine strain, we compared the major clinical parameters (arthritis onset, severity, and incidence) in commercially available BALB/c (6) and C3H colonies (7). While the intergroup (substrain) differences (50-100% incidence with arthritis scores of 5.1-12.4) occurred in BALB/c mice, the individual differences among BALB/c colonies disappeared by week 9 in PG/DDA-immunized mice (data not shown). The incidence was 100%, and the mean ± SD arthritis score in 9 different BALB/c colonies (n = 462) was 10.6 ± 3.6 at week 9 (Fig. 2B); this was not observed in PG/CFAimmunized BALB/c mice at this time point (Fig. 1). In contrast to the BALB/c strain, the extreme differences among C3H colonies, for example, between C3H/HeJ and C3H/HeJCr, 2 colonies which are otherwise derived from the same founder (7), remained significant in PG/DDA-immunized mice (compare columns 4 and 5 in Fig. 2B). Only 28% of the C3H/HeJ mice (from The Jackson Laboratory) developed arthritis, and with very low arthritis scores (1.9 ± 1.2), after up to 4 antigen injections (column 5, Fig. 2B, panel 2), whereas the incidence was 100% in C3H/HeJCr mice (from the NCI), with an arthritis score of 5.7 ± 2.3 (column 4, Fig. 2B).

Chapter 2

40

Figure 2. Comparative analysis of 9 BALB/c substrains, 3 C3H substrains, F1 hybrids of BALB/c x DBA/2 (CDF1) (see refs. 6, 7, and 8, respectively), and an additional 10 inbred strains with different H-2 haplotypes. Female mice were immunized intraperitoneally with deglycosylated human cartilage PG in DDA as shown in Fig. 1, except that all animals, regardless of their stage of arthritis, received 4 PG/DDA injections. All animals were killed on days 91-92 of immunization (4 weeks after the fourth antigen injection). Data from the end of the experiments in all animals, whether arthritic or nonarthritic, are shown (n = 6-16 animals per group). A, Inbred mouse strains that were studied. B, Incidence of arthritis, and arthritis scores. C, Antigenspecific T cell responses: proliferation (expressed as a stimulation index; S.I.), production of interferon-γ (IFNγ), and production of interleukin-4 (IL-4). D, Serum level of heteroantibodies, serum level of autoantibodies, and the IgG1:IgG2a ratios in PG/DDA-immunized mice. Significant differences between groups or pooled groups are indicated: * = P<0.05; ** = P<0.01. The corresponding major histocompatibility complex haplotype of each strain is shown across the bottom of A and B (panel 1). In each panel, the groups are shown in the same order as in A. Values are the mean and SD. See Fig. 1 for other definitions.

A

C

B

D

A

C

B

D


41

Although all BALB/c mice and the highly susceptible C3H colonies uniformly reached 100% incidence of arthritis by weeks 9-10 in response to PG/DDA immunization (Fig. 2B, panel 1), there were significant differences in the final arthritis scores in BALB/c mice (arthritis score 11.6 ± 1.6) compared with the highly susceptible C3H colonies (arthritis score 7.2 ± 2.1) (Fig. 2B). PGIA susceptibility in inbred murine strains Susceptibility to PGIA is determined by both major histocompatibility complex (MHC) and non-MHC genes (8,24), and the different colonies of susceptible BALB/c (H-2d) and C3H (H-2k) strains exhibit variabilities in both the severity and incidence of PGIA (Fig. 2B). Therefore, the next evident question was whether other inbred murine strains with the same (H-2d or H-2k) or different haplotypes can develop PGIA in response to PG/DDA immunization. While all murine strains, regardless of their H-2 haplotype or genetic background, responded well to PG/DDA immunization (Figs. 2C-D), we could not find any strain other than BALB/c and C3H that was susceptible to PGIA (Fig. 2B, panel 1). This was especially interesting because a number of murine strains carried the same class II alleles (either H-2d or H-2k haplotype) and exhibited similar, or occasionally even higher, T cell and B cell responses to PG immunization compared with the susceptible BALB/c or C3H colonies (Figs. 2C-D). Arthritogenic effect of immunization with DDA and PGs isolated from various species In previous studies, we tested cartilage PGs from various species for their ability to induce arthritis in BALB/c mice when used with Freund’s adjuvants (4,8). PGs from fetal human and pig cartilage and from adult human and dog cartilage were the only ones that could induce arthritis in BALB/c mice, but only if the glycosaminoglycan (chondroitin sulfate or both chondroitin sulfate and keratin sulfate) side chains were removed (5,6). Cartilage PGs from other species either did not induce arthritis or induced only a weak and transient inflammation at a very low incidence (<10%) in BALB/c mice immunized with PG in Freund’s adjuvants. Using DDA as adjuvant, we retested a few, relatively easily accessible cartilage PGs of various species. Quite unexpectedly, PGs from fetal and adult bovine cartilage and from adult pig and sheep cartilage proved as arthritogenic as PGs isolated from newborn or adult human cartilage when they were injected with DDA. Moreover, crude cartilage extract from osteoarthritic cartilage proved to be excellent arthritogenic material if both the chondroitin sulfate and keratin sulfate side chains of PG were removed (6); it was as effective as the highly purified human cartilage PG (Table 1). Thus, when using lipophilic adjuvant DDA for immunization, PGs with relatively poor arthritogenicity in Freund’s adjuvants can induce arthritis with maximum incidence and high severity in mice of the BALB/c strain (Table 1). While cartilage macromolecules other than PG (Table 1) in crude extracts have also provoked immune responses, these molecules of PG- and link protein-free cartilage extracts were not arthritogenic in BALB/c mice (see PG-free extract of human articular cartilage, Table 1).

Chapter 2

42

Table 1. Arthritogenic effect of different species of cartilage PGs on BALB/c mice and their immune responses to mouse cartilage PG*

Serum autoantibody to mouse PG,

mean ± SD mg/ml

Cartilage PG source (deglycosylation)

Day of earliest onset†

Arthritis incidence‡

Day of maximum incidence§

Cumulative severity score,

mean ± SD

Proliferation, S.I.,

mean ± SD¶ IgG1 IgG2a

Human newborn articular cartilage PG (ABC)

26 15/15 (100) 38 11.8 ± 3.8 3.8 ± 1.3 0.5 ± 0.1 0.1 ± 0.0

Human adult articular cartilage PG (ABC/endo-β-gal)

26 15/15 (100) 46 12.4 ± 1.5 2.7 ± 0.3 0.4 ± 0.2 0.1 ± 0.0

OA-24; human OA cartilage extract (ABC/endo-β-gal)

26 44/44 (100) 48 9.9 ± 1.7 2.4 ± 1.0 0.3 ± 0.1 0.1 ± 0.1

Human adult articular cartilage PG (ABC)

42 9/9 (100) 58 8.8 ± 3.4 2.9 ± 2.1 0.3 ± 0.2 0.1 ± 0.0

Canine articular cartilage PG (ABC) 27 15/15 (100) 47 12.8 ± 1.6 4.7 ± 3.0 0.5 ± 0.2 0.1 ± 0.0

Bovine nasal/articular cartilage PG (ABC)

47 20/20 (100) 66 8.8 ± 3.2 2.7 ± 0.1 0.1 ± 0.0 ND

Fetal bovine articular cartilage PG (ABC)

48 9/9 (100) 68 13.5 ± 2.4 2.4 ± 0.2 0.1 ± 0.0 ND

Porcine articular cartilage PG (ABC/endo-β-gal)

32 9/9 (100) 51 12.2 ± 1.8 2.8 ± 0.5 0.1 ± 0.0 0.1 ± 0.0

Porcine articular cartilage PG (ABC) 32 4/5 (80) 68 12.8 ± 2.9 2.0 ± 0.2 0.2 ± 0.0 ND

Sheep articular cartilage PG (ABC/endo-β-gal)

45 5/5 (100) 70 7.8 ± 3.0 2.1 ± 0.3 0.1 ± 0.0 ND

Chicken sternal cartilage PG (ABC) NA 0/14 (0) NA NA 5.6 ± 2.2 0.1 ± 0.0 ND

Newborn mouse PG NA 0/8 (0) NA NA 1.1 ± 0.1 0.2 ± 0.1 ND

Rat Swarm chondrosarcoma PG (ABC) NA 0/12 (0) NA NA 1.0 ± 0.2 0.7 ± 0.2 0.1 ± 0.0

PG-free human adult articular cartilage extract

NA 0/10 (0) NA NA 6.4 ± 1.4 ND ND

PBS + DDA NA 0/22 (0) NA NA 0.9 ± 0.2 ND ND

Type II collagen from human adult articular cartilage

NA 0/15 (0) NA NA 3.6 ± 1.4 ND ND

* Mice were immunized with purified cartilage proteoglycan (PG) aggrecan (100 µg of PG core protein in 100 µl of phosphate buffered saline; PBS) and dimethyldioctadecylammonium bromide (DDA; 1 mg/100 µl). The OA-24 group received crude cartilage extract from osteoarthritic human cartilage (~100 µg of PG core protein). Native (nonglycosylated) cartilage PGs did not induce arthritis when used in emulsion with either of the Freund’s adjuvants (4,8) or with DDA (results not shown). The type of deglycosylation required to retrieve the arthritogenic potential of a given cartilage PG sample is shown. Chondroitin sulfate side chains were removed by chondroitinase ABC (ABC) treatment; both chondroitin sulfate and keratan sulfate side chains were depleted by chondroitinase ABC and endo-β-galactosidase (ABC/endo-β-gal) treatment. Each animal received 4 injections, regardless of whether arthritis developed at any immunization time point. Mice were killed at the same time on days 91-92. Experiments were repeated 2 or 3 times (except those groups with 5 mice), and the results are summarized. Results for the bovine nasal and bovine articular cartilage are combined. NA = not applicable; ND = not detected. † The day when the earliest onset of arthritis (first arthritic animal) was diagnosed in the group. ‡ Number of arthritic animals/all immunized animals (%). § The day when all mice in the immunized group developed arthritis, except in negative (NA) groups or in the group labeled with the double asterisk. ¶ Antigen-specific T cell proliferation in the presence of the same antigen as used for immunization (50 µg protein/ml), expressed as the stimulation index (S.I.).

Histopathologic features of the peripheral joints and spine of PGIA-susceptible mice In the mice immunized with PG in DDA, the clinical scores for the paws corresponded well to the histopathologic abnormalities in the small peripheral joints, as described for the “classic” PG/CFA-induced form of PGIA (3,4,6-8). Joint inflammation started with mononuclear (mostly lymphocyte) and polymorphonuclear infiltration, which was soon accompanied by massive cartilage degradation, followed by bone erosion (Figs. 3B-D).


43

Figure 3. Clinical appearance and histopathologic and radiographic features of PGIA in BALB/c mice immunized with PG/DDA. A, Normal ankle joint. B, Heavily inflamed ankle joint 2-3 weeks after the onset of arthritis. Insets, Photographs of the hind paws from which the respective tissues were obtained. C and D, Higher-magnification views of tarsometatarsal joints at 2 weeks and 4 weeks, respectively, after arthritis onset. In the acute/subacute phase of the disease, the same joint may have massive lymphocyte and neutrophilic granulocyte infiltration, as well as extensive cartilage and bone erosions (B-D). The erosions are the consequences of invasion by the aggressive pannus-like synovial tissue (pannus). Shaded arrows with white arrowheads in C indicate the borders between bone and bone-resorbing soft tissue (pannus). Arrowheads in C, D, and H indicate areas of cartilage erosion. E, Normal knee joint. Jsp = joint space. F, Knee joint obtained on the day of arthritis onset, showing very early inflammation, as indicated by thickening of the synovial lining cell layer (parallel arrows). G and H, Knee joints obtained 4 days and 12 days, respectively, after arthritis onset. Arthritis onset was recorded as paw inflammation. Dotted lines in G and H and arrowheads indicate cartilage erosions. Curved white arrow in G indicates proliferation of a pannus-like granulomatous tissue at the bone–cartilage junction. I-K, Representative micrographs of intervertebral discs from a normal, nonimmunized animal (I), and from mice with PGIA (J and K). J, Histologic features of a relatively acute (early) onset of spondylitis (~2-3 weeks after the onset of inflammation in the peripheral joints) with hyperproliferative pannus-like granulation tissue (curved white arrows). This aggressive, pannus-like tissue destroys the annulus fibrous and resorbs the nucleus pulposus. K, Histologic features of a completely resorbed intervertebral disk, with segments of vertebrae ankylotized by fibrotic tissue and chondrophytes, ~8-10 weeks after the onset of peripheral joint inflammation. Insets, Radiographs of the spines from which the respective tissues were obtained. All tissue sections were stained with hematoxylin and eosin. See Fig. 1 for other definitions. While these histopathologic abnormalities were very similar to those previously described in either BALB/c or C3H mice immunized with PG in Freund’s adjuvants (3,6,7), comparison of the 2 immunization protocols revealed remarkable differences in the knee joint (Figs. 3E-H) and spine (Figs. 3I-K). For example, inflammation usually can be seen in the ipsilateral knee joint 2-3 weeks after the onset of paw inflammation (small joints) in PG/CFA-immunized BALB/c mice. In contrast, both small joints and ipsilateral large joints (Figs. 3F-G) were simultaneously inflamed in PG/DDA-immunized mice, but the periarticular inflammation (edema and cellular infiltration) was less pronounced in the large joints. A second remarkable difference was that, simultaneously with the synovial joint inflammation, massive spondylitis was detected in all arthritic BALB/c mice that had been immunized with PG/DDA

Chapter 2

44

(Figs. 3J-K). This was especially unusual, because spondylitis typically only appeared 2-4 months after peripheral (synovial) joint inflammation in BALB/c mice that had been immunized with PG/Freund’s adjuvants (3,6-8). Mechanisms of action of DDA as adjuvant As briefly summarized in Materials and Methods, a large number of preliminary and supplemental experiments (protocols 5-9) were performed to determine the optimum conditions (dose, combination, and route) of PG/DDA immunization for induction of PGIA. The overall results of T cell- and B cell-mediated immune responses to human PGs in PG/CFA- or PG/DDA-immunized BALB/c mice, however, were highly comparable, and serum levels of proinflammatory cytokines were even more pronounced in PG/CFA-immunized animals than in those injected with PG/DDA (Table 2).

Table 2. Immune responses and antigen-specific cytokine production by spleen cells after immunization of BALB/c mice with cartilage PG in CFA or in DDA*

After first injection After second injection After third injection PG in CFA PG in DDA PG in CFA PG in DDA PG in CFA PG in DDA

T cell proliferation, S.I. 2.1 ± 0.25 4.25 ± 0.63† 3.42 ± 1.06 3.21 ± 0.95 3.49 ± 1.13 3.46 ± 1.10 Serum antibody level, mg/ml

IgG1 to human PG <0.02 ND 6.62 ± 3.90 3.36 ± 1.42† 10.6 ± 0.42 9.6 ± 1.52†

IgG2a to human PG <0.02 <0.02 0.09 ± 0.04 0.06 ± 0.05 0.24 ± 0.06 0.38 ± 0.12

IgG1 to mouse PG <0.02 0.10 ± 0.01 0.28 ± 0.18 0.06 ± 0.04† 0.65 ± 0.13 0.45 ± 0.09†

IgG2a to mouse PG ND ND 0.03 ± 0.00 0.02 ± 0.01 0.03 ± 0.00 0.05 ± 0.00

Serum cytokine level, pg/ml

IL-1β 28.5 ± 12.3 10.8 ± 0.11‡ 20.7 ± 0.22 15.0 ± 2.60† 23.7 ± 5.60 19.7 ± 0.16†

TNFα 77.3 ± 28.4 47.2 ± 12.3‡ 405 ± 55.4 239 ± 23.4‡ 587 ± 44.2 441 ± 55.7†

IL-6 98.6 ± 11.3 72.9 ± 18.3‡ 101 ± 34.6 86.6 ± 19.6‡ 158 ± 28.2 118 ± 18.3‡

IL-10 ND 10.9 ± 0.09 10.1 ± 12.4 3.5 ± 0.11 4.2 ± 1.8 5.4 ± 0.08

Antigen-specific cytokine production by spleen cells, pg/108 cells/ml

IFNγ ND 55.0 ± 9.6 12.1 ± 4.3 55.0 ± 9.40 46.8 ± 12.1 107 ± 32.4

IL-4 ND ND 41.2 ± 5.8 48.2 ± 13.2 25.8 ± 9.70 34.8 ± 8.60

IFNγ:IL-4 ratio NA NA 0.29 1.14‡ 1.81 3.07‡

IL-12 ND ND ND 11.2 ± 2.60 ND 23.0 ± 11.3

TNFα ND ND 9.28 ± 0.9 1.20 ± 0.02† 10.0 ± 4.72 6.40 ± 0.91†

Ratio of lymphocytes in spleen

CD3+:B220+ NE NE NE NE 0.83 ± 0.12 0.81 ± 0.06

CD4+:CD8+ NE NE NE NE 0.95 ± 0.02 1.13 ± 0.04†

* Assays were performed 9-11 days after the intraperitoneal injection. Values are the mean ± SD. PG = proteoglycan aggrecan; CFA = Freund’s complete adjuvant; DDA = dimethyldioctadecylammonium bromide; S.I. = stimulation index; ND = not detected or the same amount was measured in nonstimulated and PG-stimulated spleen cell cultures; IL-1β = interleukin-1β; TNFα = tumor necrosis factor-α; IFNγ = interferon-γ; NA= not applicable; NE = not evaluated. † P = 0.05 versus PG in CFA-immunized group. ‡ P = 0.01 versus PG in CFA-immunized group.


45

Although some studies indicate that the use of DDA for vaccination provokes a more dominant delayed-type hypersensitivity reaction, accompanied by an antibody response that is higher than that obtained with the use of many other adjuvants, including CFA (26), we could not confirm those observations. The serum antibody levels to either PG or CII, or against a number of arthritis-irrelevant antigens (such as keyhole limpet hemocyanin, BSA, or ovalbumin; data not shown) were similar or even lower when injected with DDA than with CFA. In contrast, we found

extensive differences in antigen-specific IgG1:IgG2a and IFNγ:IL-4 ratios when the 2 adjuvants were compared

throughout the immunizations (Tables 2 and 3). Again, while the overall effects of DDA and CFA on the T cell response (measured as PG-specific T cell proliferation and IL-2 production) were highly comparable (Table 2), antigen (PG)-specific cytokine production by either PG-primed spleen cells (Table 2) or peripheral lymph node cells (data not shown) was significantly different and was clearly shifted toward Th1.

Table 3. Incidence, severity, and immune responses in DBA/1J mice immunized with human CII in CFA or DDA* ID injection ID injection IP injection

CII in CFA, males

CII in DDA, males

CII in CFA, females

CII in DDA, females

CII in CFA, males

CII in DDA, females

In vitro T cell proliferation in the presence of 25 µg/ml of human CII

Stimulation index 3.4 ± 1.0 2.9 ± 0.8 2.7 ± 1.0 2.4 ± 1.6 3.1 ± 1.6 2.3 ± 1.6

Serum anti-mouse CII antibodies, µg/ml

IgG1 407 ± 28 490 ± 133 398 ± 54 452 ± 78 426 ± 114 341 ± 222

IgG2a 48 ± 17 23 ± 14 66 ± 21 53 ± 22 19 ± 11 20 ± 9.1

Clinical symptoms

Day of earliest onset 30 30 30 30 40 32

Maximum arthritis score 5.8 ± 3.4 8.2 ± 2.2† 5.2 ± 2.4 8.3 ± 3.2† 2.4 ± 0.6 2.4 ± 1.2

Maximum incidence, % 85 100† 90 100 25‡ 25‡

* DBA/1J mice were immunized with 100 µg of human type II collagen (CII) emulsified with 100 µl of Freund’s complete adjuvant (CFA) or with 100 µl of dimethyldioctadecylammonium bromide (DDA; micelle form) and injected intradermally (ID) into the proximal tail or intraperitoneally (IP). Mouse CII was used for measuring autoantibodies; human CII was used for measuring T cell proliferation. Values are the mean ± SD of 2 independent experiments (n = 20 mice per group). † P = 0.05 versus group injected with CII in CFA. ‡ P = 0.01 versus each of the groups injected intradermally. To gain insight into the local mechanisms of DDA action and to understand why DDA supported PGIA more powerfully, BALB/c mice were injected IP with PBS, PBS/IFA, PBS/CFA, or PBS/DDA, with or without cartilage PG antigen, and cells obtained from a peritoneal lavage and from the spleen were harvested at 6 h, 12 h, every 24 h thereafter until day 9, on days 14 and 21 after the first injection, and 9 days after the second or third injection. As expected, the cell number was significantly higher, and reached a peak after 24-48 h, in the peritoneal lavage fluid from all adjuvant-injected groups compared with the PBS- or PG/PBS-injected groups, and these levels never returned to normal (data not shown). The cell influx contained predominantly neutrophilic granulocytes (up to 70-85%), and the ratio of the neutrophils was consistently highest in the CFA-injected groups at every time point evaluated. The cell number in peritoneal lavage fluid from DDA-injected groups was approximately two-thirds to one-half the cell number in fluid from CFA-injected animals by days 9-14 (data not shown). The cells consisted almost

Chapter 2

46

exclusively of F4/80+ macrophages (mean ± SD, 79 ± 11%) and CD3+ T cells (7.0 ± 2.6%), and more than 60% of the CD4+ cells were activated (CD4+/CD69+). Moreover, the ratio and the total number of CD11c+ cells (activated dendritic cells) in the peritoneal lavage fluid were highest in the PG/DDA-injected group (data not shown). Application of DDA in the induction of CIA Induction of PGIA required IP immunization of genetically susceptible strains of mice with deglycosylated cartilage PGs, and the synergistic adjuvant effect of DDA was significantly higher than the effect of even the most powerful batch of CFA. Therefore, the question of whether DDA could be used to replace CFA in other frequently used arthritis models was evident. We obtained the same results using mBSA/DDA immunization for antigen-induced arthritis (data not shown), as described previously for mBSA/CFA immunization (18). The most important data for CIA are summarized in Table 3. The onset of CIA was highly comparable in various groups of DBA/1 mice, but 100% incidence was reached only in CII/DDA-immunized mice after 2 antigen injections, with a significantly higher arthritis score in CII/DDA-immunized mice than in CII/CFA-immunized mice (Table 3). The route of immunization (ID versus IP), however, was a critical component of arthritis induction (Table 3), and the incidence reached only 60% even after the third IP injection of CII given either in CFA or DDA (data not shown). It is important to note here that although both murine strains (BALB/c and DBA/1) responded well to immunization with either human PG or CII, DBA/1 mice did not develop arthritis in response to PG/DDA immunization (data not shown), and BALB/c mice did not show signs of arthritis after CII/DDA (Table 1) or CII/CFA immunization (6).


47

Discussion We present herein a summary of the results of our studies using a powerful adjuvant DDA for immunization and arthritis induction. DDA belongs to the group of lipophilic quaternary amines that were described more than 35 years ago as a potential adjuvant (27). DDA as an adjuvant has been used successfully and without side effects in vaccines administered to children and pregnant women (28,29) It is also widely used as a detergent in cosmetic compounds and fabric softeners (for review, see ref. 30). It is a positively charged compound (MW 631 daltons) with a monovalent counter-ion (Br), and 2 long alkyl chains that are hydrophobic (27,30). Because of the lipophilic chains, DDA is poorly soluble in water but forms a semicolloidal polycationic liposomal micelle structure at a temperature of 40°C or higher (22,30). DDA is a highly potent immunostimulator, especially with negatively charged antigens, provoking a strong delayed-type hypersensitivity (30-33). It is a powerful, nonirritant adjuvant and, via T cell stimulation, significantly enhances antigenspecific B cell activation and immunoglobulin production (26,34). A special benefit of the use of DDA in rodent models of autoimmunity is that this adjuvant forces the immunoregulation toward Th1 (35-37). DDA can enhance the adjuvant effect of other adjuvants (14) and even potentiate the arthritogenic effect of IFA (14) or CFA (38) in oil adjuvant-susceptible strains of rats. However, DDA as an adjuvant has never previously been tested in any autoimmune model. We compared the effects of DDA with those of CFA and IFA in arthritis-susceptible murine strains to gain insight as to how the DDA could achieve an even more superior effect than the Freund’s adjuvants. While the overall immune responses (antibody production and antigen-specific T cell responses) were highly comparable when human PGs in CFA or in DDA were used, suboptimal doses of cartilage PGs or PGs having only a suboptimal arthritogenic effect when injected with CFA were as effective as the human cartilage PGs in provoking PGIA. Moreover, crude extracts obtained from osteoarthritic cartilage, when appropriately deglycosylated, induced arthritis in a manner similar to that of the highly purified human cartilage PG, whereas other cartilage matrix macromolecules in the crude extract did not induce arthritis in BALB/c mice. There are a number of inducible animal models of human autoimmune diseases, all of which require immunization of genetically susceptible strains of rodents and non-human primates with a target organ-specific antigen in adjuvant (15,39,40). Depending on the model and species, 1-4 injections of antigen are required, and at least 1 of these injections should be given with CFA as the adjuvant (15). Therefore, CFA seems to be a critical component in the induction and achievement of a high incidence and severity of autoimmune diseases in rodent models. The use of the same (auto)antigen in IFA, Alhydrogel (aluminum hydroxide gel), or synthetic adjuvants is either insufficient to induce the disease, or the incidence and severity are far below those of the antigen/CFA-induced disease. Thus, the nonspecific activation of innate immunity (macro-phages, dendritic cells) and T cells by mycobacterial components such as muramyldipeptide (peptidoglycan), HSP, and trehalosedimycolate (a glycolipid equivalent with lipopolysaccharide of Escherichia coli), in mineral oil (14,15) is a critical component in the provocation of immune reactions to self antigens in autoimmune models. While a large number of adjuvants can be used for immunization of animals or for vaccination of humans, until now CFA has remained the most potent and powerful adjuvant in all experimental models of autoimmunity. The use of CFA, however, induces several side effects, including immune reactions to mycobacterial components, and, as a

Chapter 2

48

highly irritating compound, CFA induces sterile inflammation, followed by local granulomatous tissue formation and severe adhesions, especially in the peritoneal cavity (15). In our comparisons of the 2 adjuvants (CFA and DDA), only a few notable differences between the PG/CFA-immunized mice and the PG/DDA-immunized mice were noted. These were as follows: 1) significantly higher levels of serum proinflammatory cytokine and PG-specific IgG1, but not IgG2a, in PG/CFA-injected mice (Table 2); 2)

significantly higher ratios of antigenspecific IFNγ to IL-4 and significantly lower ratios of IgG1 to IgG2a in PG/DDA-

immunized mice than in PG/CFA-immunized mice; 3) the highest ratio of activated CD3+,CD44High cells in PG/DDA-primed lymph nodes; 4) an increased ratio of CD4+ to CD8+ cells in the spleens of arthritic mice (Table 2) but not in PG/DDA-primed lymph nodes; and 5) a significantly increased number of CD4+/CD69+ cells in peritoneal lavage fluid from PG/DDA-primed than in that from PG/CFA-primed BALB/c mice. Moreover, at least a 2-4-fold increase in macrophage influx into the peritoneal cavity, accompanied by more CD11c+ dendritic cells but significantly fewer polymorphonuclear cells, was characteristically observed in DDA-injected mice. Among the adjuvants tested to date, only CFA and DDA proved to express sufficient power to provoke arthritis in PG- or CII-immunized genetically susceptible strains of mice. In terms of humoral and cellular immune responses to PG or other antigens and production of related cytokines, however, CFA is at least an equivalent, and frequently, an even more effective, adjuvant than DDA (Tables 2 and 3), and both CFA and DDA induced significantly higher immune responses and cytokine production than did PG/IFA. However, the production of antigen-specific Th1 and Th2 cytokines and the shift of the Th1/Th2 balance toward Th1 were significantly more pronounced in animals immunized with PG/DDA than in those immunized with PG/Freund’s adjuvants (Table 2). As a result, the use of DDA accelerated the development of a more severe arthritis, and suboptimum doses of PG or CII antigens were able to induce inflammation via a more potent activation of the innate immunity. A potential or critical role of innate immunity in arthritis induction is consistent with the findings of studies that used only adjuvants (nonspecific stimulators) in genetically susceptible strains of rodents (14,15). Observations from these studies of oil-, pristane-, and squalene-induced arthritis (13-15) together with our observations raise the question of whether the “adjuvants” can also play a role in, and/or contribute to, joint inflammation in genetically susceptible humans. Can immunostimulatory molecules from microbes, environmental compounds (cosmetics or laundry detergents), or endogenous “self” adjuvants (such as lipid squalene) in fact cause or contribute to joint inflammation? Since a number of potential autoantigens (type II collagen, proteoglycan aggrecan, link protein, gp-39, or glucosephosphate isomerase) can be identified in various subsets of RA patients, it may be an attractive hypothesis that a nonspecific stimulation of innate immunity might be an initial component in the disease mechanism. A defect in immune regulation, the production of cytokines/chemokines, and the possible involvement of joint-derived or joint-independent autoantigens (such as rheumatoid factor) may be only subsequent, although clearly detectable, events in an initial nonspecific activation of innate immunity. While the relevance of this hypothesis requires extensive studies, our observations using an innocuous component as adjuvant seem to support this possibility.


49

Acknowledgments The authors acknowledge the contributions of many colleagues of the Section of Biochemistry and Molecular Biology, Rush University, to this work. In particular, the authors thank Dr. Alison Finnegan for her valuable comments, and Dr. Zoltán Szabó, Dr. Thomas Koreny, Sonja Velins, and Tiffany Lamb for their assistance. This work was supported in part by NIH grants (AR-40310, AR-45652 and AR-51163) and a grant from NWO (016.026.010) (SB, CB).

Chapter 2

50

References (1) Trentham DE, Townes AS, Kang AH. Autoimmunity to type II collagen an experimental model of arthritis. J Exp Med

1977;146:857-68. (2) Courtenay JS, Dallman MJ, Dayan AD, Martin A, Mosedale B. Immunisation against heterologous type II collagen

induces arthritis in mice. Nature 1980;283:666-8. (3) Glant TT, Mikecz K, Arzoumanian A, Poole AR. Proteoglycan-induced arthritis in BALB/c mice. Clinical features and

histopathology. Arthritis Rheum 1987;30: 201-12. (4) Mikecz K, Glant TT, Poole AR. Immunity to cartilage proteoglycans in BALB/c mice with progressive polyarthritis and

ankylosing spondylitis induced by injection of human cartilage proteoglycan. Arthritis Rheum 1987;30:306-18. (5) Glant TT, Buzas EI, Finnegan A, Negroiu G, Cs-Szabo G, Mikecz K. Critical roles of glycosaminoglycan side chains of

cartilage proteoglycan (aggrecan) in antigen recognition and presentation. J Immunol 1998;160: 3812-9. (6) Glant TT, Cs-Szabo G, Nagase H, Jacobs JJ, Mikecz K. Progressive polyarthritis induced in BALB/c mice by aggrecan

from normal and osteoarthritic human cartilage. Arthritis Rheum 1998;41:1007-18. (7) Glant TT, Bardos T, Vermes C, Chandrasekaran R, Valdez JC, Otto JM, Gerard D, Velins S, Lovasz G, Zhang J,

Mikecz K, Finnegan A. Variations in susceptibility to proteoglycan-induced arthritis and spondylitis among C3H substrains of mice: evidence of genetically acquired resistance to autoimmune disease. Arthritis Rheum 2001;44:682-92.

(8) Glant TT, Finnegan A, Mikecz K. Proteoglycan-induced arthritis: immune regulation, cellular mechanisms, and genetics. Crit Rev Immunol 2003;23:199-250.

(9) Pearson CM. Development of arthritis, periarthritis and periostitis in rats given adjuvant. Proc Soc Exp Biol Med 1956;91:95-101.

(10) Anderton SM, van der Zee R, Prakken B, Noordzij A, van Eden W. Activation of T cells recognizing self 60-kD heat shock protein can protect against experimental arthritis. J Exp Med 1995;181:943-52.

(11) Wooley PH, Sud S, Whalen JD, Nasser S. Pristane-induced arthritis in mice. V. Susceptibility to pristane-induced arthritis is determined by the genetic regulation of the T cell repertoire. Arthritis Rheum 1998;41:2022-31.

(12) Holmdahl R, Kvick C. Vaccination and genetic experiments demonstrate that adjuvant-oil-induced arthritis and homologous type II collagen-induced arthritis in the same rat strain are different diseases. Clin Exp Immunol 1992;88:96-100.

(13) Lorentzen JC, Klareskog L. Susceptibility of DA rats to arthritis induced with adjuvant oil or rat collagen is determined by genes both within and outside the major histocompatibility complex. Scand J Immunol 1996;44:592-8.

(14) Lorentzen JC. Identification of arthritogenic adjuvants of self and foreign origin. Scand J Immunol 1999;49:45-50. (15) Billiau A, Matthys P. Modes of action of Freund's adjuvants in experimental models of autoimmune diseases. J Leukoc

Biol 2001;70:849-60. (16) Glant TT, Mikecz K, Poole AR. Monoclonal antibodies to different protein-related epitopes of human articular cartilage

proteoglycans. Biochem J 1986;234:31-41. (17) Antonopoulos CA, Axelsson I, Heinegard D, Gardell S. Extraction and purification of proteoglycans from various types

of connective tissue. Biochim Biophys Acta 1974;338:108-19. (18) Bardos T, Kamath RV, Mikecz K, Glant TT. Anti-inflammatory and chondroprotective effect of TSG-6 (tumor necrosis

factor-alpha-stimulated gene-6) in murine models of experimental arthritis. Am J Pathol 2001;159:1711-21.


51

(19) Czipri M, Otto JM, Cs-Szabo G, Kamath RV, Vermes C, Firneisz G, Kolman KJ, Watanabe H, Li Y, Roughley PJ, Yamada Y, Olsen BR, Glant TT. Genetic rescue of chondrodysplasia and the perinatal lethal effect of cartilage link protein deficiency. J Biol Chem 2003;278:39214-23.

(20) Miller EJ. Structural studies on cartilage collagen employing limited cleavage and solubilization with pepsin. Biochemistry (Mosc) 1972;11:4903-9.

(21) Mikecz K, Brennan FR, Kim JH, Glant TT. Anti-CD44 treatment abrogates tissue oedema and leukocyte infiltration in murine arthritis. Nat Med 1995;1:558-63.

(22) Tsuruta LR, Quintilio W, Costa MH, Carmona-Ribeiro AM. Interactions between cationic liposomes and an antigenic protein: the physical chemistry of the immunoadjuvant action. J Lipid Res 1997;38:2003-11.

(23) Bardos T, Mikecz K, Finnegan A, Zhang J, Glant TT. T and B cell recovery in arthritis adoptively transferred to SCID mice: antigen-specific activation is required for restoration of autopathogenic CD4+ Th1 cells in a syngeneic system. J Immunol 2002;168:6013-21.

(24) Adarichev VA, Valdez JC, Bardos T, Finnegan A, Mikecz K, Glant TT. Combined autoimmune models of arthritis reveal shared and independent qualitative (binary) and quantitative trait loci. J Immunol 2003;170:2283-92.

(25) Buzas EI, Hollo K, Garzo M, Lukacz K, Glant TT. Lack of evidence for molecular mimicry between 65kD heat shock protein (HSP) and cartilage proteoglycan (PG), which could provoke autoimmune responses and arthritis. Arthritis Rheum 1995;38:S295.

(26) Gordon WC, Prager MD, Carroll MC. The enhancement of humoral and cellular immune responses by dimethyldioctadecylammonium bromide. Cell Immunol 1980;49:329-40.

(27) Gall D. The adjuvant activity of aliphatic nitrogenous bases. Immunology 1966;11:369-86. (28) Stanfield JP, Gall D, Bracken PM. Single-dose antenatal tetanus immunisation. Lancet 1973;1:215-9. (29) Chambers JD, Thomas CR, Hobbs JR. Induction of specific transplantation tolerance in man by autoblast

immunisation. Blut 1980;41:229-36. (30) Hilgers LA, Snippe H. DDA as an immunological adjuvant. Res Immunol 1992;143:494-503. (31) Snippe H, Johannesen L, Lizzio E, Merchant B. Variable expression of delayed hypersensitivity in different mouse

strains using dimethyl dioctadecyl ammonium bromide as an adjuvant. Immunology 1980;39:399-405. (32) Baechtel FS, Prager MD. Interaction of antigens with dimethyldioctadecylammonium bromide, a chemically defined

biological response modifier. Cancer Res 1982;42:4959-63. (33) Kraaijeveld CA, la Riviere G, Benaissa-Trouw BJ, Jansen J, Harmsen T, Snippe H. Effect of the adjuvant dimethyl

dioctadecyl ammonium bromide on the humoral and cellular immune responses to encephalomyocarditis virus. Antiviral Res 1983;3:137-49.

(34) Hilgers LA, Snippe H, Jansze M, Willers JM. Synthetic sulpholipopolysaccharides: novel adjuvants for humoral immune responses. Immunology 1987;60:141-6.

(35) Katz D, Lehrer S, Galan O, Lahmi B, Cohen S. Adjuvant effects of dimethyl dioctadecyl ammonium bromide, complete Treund's adjuvant and aluminium hydroxide on neutralizing antibody, antibody-isotype and delayed-type hypersensitivity responses to Semliki Forest virus in mice. FEMS Microbiol Immunol 1991;76:305-20.

(36) Leal IS, Smedegard B, Andersen P, Appelberg R. Failure to induce enhanced protection against tuberculosis by increasing T-cell-dependent interferon-gamma generation. Immunology 2001;104:157-61.

(37) Klinguer-Hamour C, Libon C, Plotnicky-Gilquin H, Bussat MC, Revy L, Nguyen T, Bonnefoy JY, Corvaia N, Beck A. DDA adjuvant induces a mixed Th1/Th2 immune response when associated with BBG2Na, a respiratory syncytial virus potential vaccine. Vaccine 2002;20:2743-51.

Chapter 2

52

(38) Mia MY, Zhang L, Hossain A, Zheng CL, Tokunaga O, Kohashi O. Dimethyl dioctadecyl ammonium bromide (DDA)-induced arthritis in rats: a model of experimental arthritis [In Process Citation]. J Autoimmun 2000;14:303-10.

(39) Swanborg RH. Experimental autoimmune encephalomyelitis in the rat: lessons in T-cell immunology and autoreactivity. Immunol Rev 2001;184:129-35.

(40) t'Hart BA, Vervoordeldonk M, Heeney JL, Tak PP. Gene therapy in nonhuman primate models of human autoimmune disease. Gene Ther 2003;10:890-901.

chapter 2 achievement of a synergistic adjuvant effect on ... · chapter 2 achievement of a...

Documents